diff options
author | Usama Arif <usama.arif@arm.com> | 2019-05-23 14:20:33 +0100 |
---|---|---|
committer | Usama Arif <usama.arif@arm.com> | 2019-05-24 16:11:16 +0000 |
commit | 0a5a57a3f794de851408bae1c63b1660b4c5cbe7 (patch) | |
tree | 5d49d85165a3d9fd257c5676691fc5ae24fda4f3 /arm_compute/core/NEON/wrapper/intrinsics | |
parent | 43c1bf8e47098d6ab07be5b5e685bda6d8b80344 (diff) | |
download | ComputeLibrary-0a5a57a3f794de851408bae1c63b1660b4c5cbe7.tar.gz |
COMPMID-2160: Implement Round for NEON
Change-Id: Ie80e2ad294eaf95bc823d979842c320e8fb41f67
Signed-off-by: Usama Arif <usama.arif@arm.com>
Reviewed-on: https://review.mlplatform.org/c/1215
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Reviewed-by: Georgios Pinitas <georgios.pinitas@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'arm_compute/core/NEON/wrapper/intrinsics')
-rw-r--r-- | arm_compute/core/NEON/wrapper/intrinsics/intrinsics.h | 1 | ||||
-rw-r--r-- | arm_compute/core/NEON/wrapper/intrinsics/round.h | 48 |
2 files changed, 49 insertions, 0 deletions
diff --git a/arm_compute/core/NEON/wrapper/intrinsics/intrinsics.h b/arm_compute/core/NEON/wrapper/intrinsics/intrinsics.h index 012f6868d1..c9dbb2fa81 100644 --- a/arm_compute/core/NEON/wrapper/intrinsics/intrinsics.h +++ b/arm_compute/core/NEON/wrapper/intrinsics/intrinsics.h @@ -57,6 +57,7 @@ #include "arm_compute/core/NEON/wrapper/intrinsics/pmin.h" #include "arm_compute/core/NEON/wrapper/intrinsics/pow.h" #include "arm_compute/core/NEON/wrapper/intrinsics/rev64.h" +#include "arm_compute/core/NEON/wrapper/intrinsics/round.h" #include "arm_compute/core/NEON/wrapper/intrinsics/setlane.h" #include "arm_compute/core/NEON/wrapper/intrinsics/store.h" #include "arm_compute/core/NEON/wrapper/intrinsics/sub.h" diff --git a/arm_compute/core/NEON/wrapper/intrinsics/round.h b/arm_compute/core/NEON/wrapper/intrinsics/round.h new file mode 100644 index 0000000000..da63bf6cb2 --- /dev/null +++ b/arm_compute/core/NEON/wrapper/intrinsics/round.h @@ -0,0 +1,48 @@ +/* + * Copyright (c) 2019 ARM Limited. + * + * SPDX-License-Identifier: MIT + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in all + * copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE + * SOFTWARE. + */ +#ifndef __ARM_COMPUTE_WRAPPER_ROUND_H__ +#define __ARM_COMPUTE_WRAPPER_ROUND_H__ + +#include "arm_compute/core/NEON/NEMath.h" +#include <arm_neon.h> + +namespace arm_compute +{ +namespace wrapper +{ +#define VROUNDQ_IMPL(vtype, postfix) \ + inline vtype vround(const vtype &a) \ + { \ + return vroundq_rte_##postfix(a); \ + } + +VROUNDQ_IMPL(float32x4_t, f32) +#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC +VROUNDQ_IMPL(float16x8_t, f16) +#endif // __ARM_FEATURE_FP16_VECTOR_ARITHMETIC +#undef VROUNDQ_IMPL + +} // namespace wrapper +} // namespace arm_compute +#endif /* __ARM_COMPUTE_WRAPPER_ROUND_H__ */ |